Generating spatial data for marine conservation and management
- First Online:
- Cite this article as:
- Aylesworth, L., Phoonsawat, R., Suvanachai, P. et al. Biodivers Conserv (2017) 26: 383. doi:10.1007/s10531-016-1248-x
- 223 Downloads
Do fishers know best when it comes to identifying areas with rare and depleted fish species? The global conservation crisis demands that managers marshal all available datasets to inform conservation management plans for depleted species, yet the level of trust placed in local knowledge remains uncertain. This study compares four methods for inferring species distributions of an internationally traded, rare and depleted genus of marine fishes (Hippocampus spp.): the use of (i) fisher interviews; (ii) government research trawls, (iii) scientific diving surveys, and (iv) citizen science contributions. We analyzed these four datasets at the genus and individual species levels to evaluate our conclusions about seahorse spatial occurrence, diversity of species present and the cost effectiveness of sampling effort. We found that fisher knowledge provided more information on our data-poor fish genus at larger spatial scales, with less effort, and for a cheaper price than all other datasets. One drawback was that fishers were unable to provide data down to the species level. People embarking on conservation endeavors for data-poor species may wish to begin with fisher interviews and use these to inform the application of government research, scientific diving, or citizen science programs.
KeywordsCitizen science Data-poor Hippocampus Local knowledge Scientific surveys Thailand
Spatial data lies at the heart of current conservation and management efforts. Numerous management plans for species of conservation concern call for monitoring of animals or for protecting their critical habitats (NOAA 2014). Identification of important populations (Thornton and Scheer 2012) and locations with extractive activities (Reid 2007) are common research and management objectives both on land and in the ocean. The data requirements underlying these objectives are spatial in nature and are critical to achieving conservation outcomes.
Challenges remain about the best ways to generate spatial data to achieve conservation outcomes, especially for species where baseline information on local distribution is non-existent or very limited (data-poor). Common methods to generate spatial data for such species include field surveys to detect presence or absence (Mackenzie et al. 2002), camera trapping (Bender et al. 2014), acoustic sampling (Koslow 2009), and animal tagging (Lewis et al. 2009). Growing in use are citizen science programs, which encourage the public to report their observations of wildlife (Braschler 2009). Biases and benefits are associated with each of these methods (Dickinson et al. 2010; Katsanevakis et al. 2012), and guidance is still needed on where natural resource managers should prioritize scarce resources.
Considerable information about data-poor species often exists in local knowledge (Gilchrist et al. 2005), but its application for management remains uncertain (Mallory et al. 2003). Local ecological knowledge (LEK) refers to ‘knowledge generated and reproduced through human-environment interactions in specific locations by local inhabitants’ (Byg et al. 2012). LEK has supplied sightings data for rare species (Turvey et al. 2015) both on land and in the ocean, in a cost effective manner (Thornton and Scheer 2012; Turvey et al. 2014). Providing a source for historical baseline information (McClanahan et al. 2012), LEK can complement scientific data to expand spatial or temporal scales (Tobias 2010). Despite a growing movement to verify LEK (Beaudreau and Levin 2014; Shepperson et al. 2014; Turner et al. 2015), this practice remains uncommon, lending uncertainty to its use (Usher 2000). There are many biases associated with LEK, such as observation error, perception bias, shifting baselines, and telling researchers what they want to hear (Stake et al. 2005; Byg et al. 2012). This knowledge is also limited to locals’ experience (what they need to know) (Mallory et al. 2003) and tends to incorporate more than one time period (Shepperson et al. 2014). These biases influence the validity of extracting and translating LEK into scientific formats (Berkes 2009), but all forms of information carry uncertainty.
Despite the challenges of generating spatial data for marine species, current ocean management strategies are commonly spatial. Many national and international policies are geared towards creating marine protected areas (Chape et al. 2005), and many countries are now prioritizing marine spatial planning (Douvere 2008; Stelzenm et al. 2012). Common fisheries management strategies include gear restrictions in certain areas or seasonal closures (Cadrin et al. 1995; Panjarat and Bennett 2012). In order to evaluate the effectiveness of conservation and management strategies, spatial data are required on species (Moller and Berkes 2004), habitats (Chape et al. 2005), and human activities (Reid 2007).
Many international environmental treaties, such as those on global fisheries or air pollution, use spatial data as part of the monitoring and enforcement process (Lorenz 1995; Lorenzen et al. 2010), and one new application for spatial data in the ocean is with the Convention on International Trade of Endangered Species (CITES) (Rosser and Haywood 2002). The CITES treaty regulates international wildlife trade to ensure it does not contribute to the extinction of species in the trade (Vincent et al. 2013). Internationally traded wildlife species may be listed on one of three treaty Appendices, the second of which (Appendix II) requires countries to regulate trade to ensure exports are not harmful to wild populations (UNEP-WCMC 2012). As part of the enforcement process for Appendix II species, any country may be asked to provide evidence that exports are not harmful to wild populations (non-detriment finding) and specimens are legally sourced. Meeting these conditions requires spatial data in several contexts including species locations, habitats, extractive activities and management areas, although most guidance does not highlight the importance of spatial data to this process (Rosser and Haywood 2002; CITES 2013). Recently guidelines for making non-detriment findings were published for seahorses and sharks (CITES 2013; Mundy-Taylor et al. 2014), but these were not available at the time when several countries were asked to undergo the first enforcement process for marine fish in 2012 (UNEP-WCMC 2012).
The need for guidelines became apparent when Thailand (and Vietnam) were asked to make changes in their process to evaluate the sustainability of trade for marine fishes (UNEP-WCMC 2012). Thailand is the world’s largest exporter of seahorses, (Hippocampus spp.), a cryptic marine fish genus listed on the CITES Appendix II (UNEP-WCMC 2012). Seahorses were the first marine fishes to be listed on CITES Appendix II since its inception, and the first marine fishes on the CITES treaty to have countries undergo the Appendix II enforcement process (Vincent et al. 2013). Insufficient data on seahorse distribution and extractive activities were available when Thailand was asked to make changes to its process, which necessitated the exploration of appropriate data sources (CITES 2013).
This paper compares information gained and costs associated with different types of information that can be used to infer species distributions. We determined the most cost-effective method to identify where seahorses live and evaluate these methods at (i) the genus and (ii) individual species levels. We compare knowledge of these marine fishes generated from fisher interviews to data gathered from governmental research trawls, scientific diving surveys and citizen scientists contributions, and evaluate the conclusions drawn from the four methods. Two years of collaborative research efforts yielded four datasets with spatial information about seahorses (Hippocampus spp.): (i) fisher interviews, (ii) Thailand Department of Fishery (DoF) research trawls, (iii) scientific diver surveys, and (iv) citizen science diver contributions) with spatial information about seahorses (Hippocampus spp.). We determine which dataset (or combination of datasets) was the most useful in a conservation context and compared the knowledge from the various datasets in terms of in situ seahorse occurrence, species composition, and cost of generation.
We selected data from Thailand’s western coast, an area covering 865 km in length, with six provinces (Lymer et al. 2010). We chose this coast of Thailand because it had the most comprehensive coverage across all four datasets. We now describe the methods used to produce these datasets.
Dataset 1: local knowledge—fisher interviews
To select respondents, we targeted two sets of landing locations along the Andaman Coast: fishing ports for commercial fishers and coastal villages for small-scale fishers (Online Resource 2). We visited 26 locations including 87% of commercial ports (n = 7 of 8) and 3% of small-scale fisher villages (n = 19 of 621) based on the recommendations of provincial DoF staff. These ports and villages were representative of the fishing gears used in each province. For the percentage of ports and villages visited by province, see Online Resource 2.
To gather seahorse spatial data, we interviewed 73 commercial captains and 120 small-scale fishers at the ports and villages in six provinces along the Andaman Coast with semi-structured interviews. We determined the number of fishers to interview in each location as either 10% of the estimated total number of fishing boats landing catch at that location or the saturation method (Tobias 2010). We defined saturation as the point where if the 6th, 7th and 8th interviews yielded no new information on gear type, fishing grounds, or seahorse locations, then saturation had been reached (Tobias 2010). We additionally confirmed saturation through our observations of gear type at each landing location. To participate in our study, fishers had to be a captain and participated in fishing activity within the last year because we assumed that captains would have the most spatial knowledge of seascape based on navigation experience. Interviews occurred in Thai through the use of a local translator. All interviews followed UBC Human Ethics Protocols (H12-02731).
To identify seahorse locations across the Andaman Sea, we conducted spatial mapping with fishers using a computer tablet with iGIS software (Geometry Pty Ltd 2008). We oriented fishers on how to use the tablet before mapping started. If fishers stated they did not feel comfortable using the tablet after the orientation, we guided mapping efforts by drawing seahorse polygons in iGIS based on the fisher’s description. Fishers were asked to identify (a) the spatial extent of their fishing grounds (b) locations where they knew seahorses lived, and (c) places where they had captured seahorses with fishing gear. They were asked to describe the depths of all the aforementioned locations. Fishers were also asked to describe the seahorse species they found based on a seahorse species identification guide. If fishers could not identify the species, they were invited to classify the type of seahorse as either smooth or spiny.
Dataset 2: DoF research trawls
Trawl surveys were executed by the Thailand DoF in 2010, 2012 and 2013 at 22 pre-determined sampling locations throughout the Andaman Sea, originally intended for the purpose of sampling commercially exploited fish species. Each location represented a grid equivalent to 15 × 15 nautical miles (Online Resource 3). The grids were sampled four times per year using a 23.5 m otter-board trawl research vessel, with trawl speed set at 2.5 nautical miles/h. Trawling took place within each gridded area for 1 h. GPS locations were marked at the beginning and ending of each trawl, and the depth of the trawl recorded. All fish from each trawl were sorted and identified by DoF staff. Presence or absence of seahorses was noted for each trawl, and seahorses were identified to species level.
Dataset 3: scientific diving surveys
We used data from Aylesworth et al. (2015) to identify locations of seahorses in the Andaman Sea. These data were collected through underwater scientific diving surveys over two three-month field seasons. A total of 26 sites were visited over the two field seasons in coral, seagrass, mangrove and sandy soft bottom habitats (Aylesworth et al. 2015). These data included GPS coordinates for each site, and when seahorses were observed, the species, depth, and habitat were recorded (Aylesworth et al. 2015). This research was conducted in accordance with UBC Animal Ethics protocols (A12-0288).
Dataset 4: citizen science diver contributions
We evaluated citizen science contributions of seahorse sightings from the global seahorse database, iSeahorse.org. All data available through April 23, 2015 were included for analysis, excluding any data contributed by the authors. The iSeahorse program encourages divers to report sightings of seahorses to the database with the option to include species names, habitat, depth, and any additional information, e.g., behavior. A concerted effort was made during our project to promote the iSeahorse program within Thailand’s diving community through presentations, training workshops, and appeals to divers (English and Thai) on social media. We had no control over who contributed sightings and from which locations, but we used all sightings spatially located in Thai waters.
We conducted three comparisons (Fig. 1) to assess if the conclusions we drew about the locations of seahorses at the genus and individual species levels changed based on the data we used. For each analysis we compared the information on where seahorses lived, how many species were identified and which types, and the effort needed to generate the data. We included all available data from all datasets, after an initial review determined all spatial data was located inside Thailand’s exclusive economic zone. To compare the knowledge about where seahorses lived, we analyzed the total number of seahorses, depth ranges, and the number of seahorses found inshore (within 3 km of shore) compared to offshore (more than 3 km from shore). We defined inshore areas to be within 3 km of the shore, because this coastal area had been set-aside for small-scale fishers for many years (Lymer et al. 2010). Pushnets and trawlers are not allowed within 3 km of shore based on Ministerial Notification from 20 July 1972. To compare knowledge about seahorse species, we analyzed the number of species found (or reportedly found), the composition of those species, and determined the dominant species present. To compare the effort expended to generate the datasets we compared the number of sites visited, total number of days, number of interviews conducted, number of staff deployed, and cost per day in US dollars to generate each dataset. We determined cost based on the amount of money expended in Thailand to generate each dataset (Online Resource 4). For example, with the citizen science dataset, we did not include creation or maintenance costs associated with the iSeahorse.org database of seahorse sightings. We only included outreach costs conducted in Thailand from January 2013 until September 2014.
Comparison 1: commercial and small-scale fishers
To compare the identified spatial extent of seahorses across the Andaman coast from commercial and small-scale fishers, we created seahorse presence maps by overlaying individual fisher’s maps of seahorse locations. Prior to analysis, we standardized the maps’ spatial accuracy because fishers’ knowledge varied in how they described the locations of seahorses and their fishing grounds (Online Resource 5). All fisher shapefiles required some standardization to ensure spatial accuracy. By reviewing the data for spatial accuracy we also ensured there were no outliers in these data (Online Resource 5). We counted the number of shapefiles that overlapped in each area. Then we calculated the proportion of fishers interviewed who reported catching seahorses in any given area. We binned the output count shapefiles to represent equal thirds in the overlap values [Commercial (0–9, 10–19, 20–33%; Small-scale range 0–5, 5–10, 10–15%)]. We verified these maps with two methods, internal consistency and fisher interviews prior to use in analysis (Online Resource 5). A table reporting the errors identified during the verification process can be found in Online Resource 5. All spatial analyses were conducted in ArcGIS 10.1 (ESRI 2011).
Comparison 2: trawl captain interviews and DoF research trawls
We compared commercial trawl captains’ knowledge about seahorse locations with the DoF research trawl data. From the commercial fisher interviews, we selected those shapefiles from trawl captains and executed the same analysis as described above to create a seahorse presence map. With the DoF trawl survey data, we created point shapefiles. Next we executed a simple overlay analysis with the trawl captain seahorse maps and the DoF trawl survey data to identify the similarities and differences between the two datasets. We ran an additional analysis of trawling effort to highlight discrepancies in the datasets that were related to effort (Online Resource 6).
Comparison 3: small-scale fisher and diver (scientific and citizen science) generated data
We compared the knowledge generated from small-scale fishers to that of scientific diving surveys and citizen science diver contributions of seahorses throughout the Andaman Sea. We executed a simple overlay analysis with these datasets to identify spatial similarities and differences among the datasets.
Comparison 1: commercial and small-scale fishers
In terms of sampling effort, our coverage of landing locations and number of respondents varied greatly between small-scale and commercial fishers (Table 2). We visited a larger proportion of commercial landing locations than small-scale fisher landing locations, given similar metrics of cost per day, total number of days and staff needed for conducting fisher interviews (Table 2). However, we interviewed many more small-scale fishers than commercial fishers over a relatively similar time period (Table 2).
Comparison 2: trawl captain interviews and DoF research trawls
Seahorse species knowledge gathered from (1) fisher interviews, (2) DoF research trawls, (3) scientific diving surveys, and (4) citizen science contributions
Depth in meters (mean)
Majority reports inshore, offshore or approximately equal
Seahorse species (number)
DoF research trawls
H. trimaculatus (524)
H. spinosissimus (13)
H. kelloggi (1)
Scientific diving surveys
H. comes (31)
H. kuda (17)
H. spinosissimus (4)
H. trimaculatus (2)
H. kelloggi (1)
H. mohnikei (1)
Citizen science contributions
H. comes (61)
H. kuda (9)
H. spinosissimus (8)
H. histrix (6)
H. kelloggi (4)
H. trimaculatus (3)
H. mohnikei (2)
Hippocampus spp. (2)
At the individual species level, only the DoF research trawls provided information on the number of species observed, and species composition (Table 1). Hippocampus trimaculatus Leach 1814 represented 97% of the individuals captured, and approximately 50% of all seahorses were captured just south of Koh Lanta (Fig. 3). H. trimaculatus was found mostly in areas south of Phuket (Fig. 3), whereas three species were found from Phuket north to Ranong.
Sampling effort expended to generate the four spatial datasets on the Andaman coast, Thailand
# Sites visited
Total landing or sampling sites
Total # of seahorses found
# of staff deployed
Amount spent per day (USD)
Total cost for data generation (USD)
DoF research trawls
Citizen science contributions
Comparison 3: small-scale fisher and diver (scientific and citizen science) generated data
Our results emphasize the most expeditious way to generate spatial data for data-poor marine fishes is through interviews with marine stakeholders (Johannes 2000; Gilchrist et al. 2005). As with our fisher interviews for seahorses on the Andaman coast of Thailand, spatial data on rare or depleted marine fishes often exists in local knowledge (Gilchrist et al. 2005; Thornton and Scheer 2012). When urgency arises to ensure the sustainability of data-poor species, as in our study, local knowledge is cheaper to generate and provides a starting point to make inferences about species distribution at larger spatial scales (Moller and Berkes 2004; Thornton and Scheer 2012). Similar to other studies, we found supporting evidence that local knowledge was complementary to scientific data (Castellanos-Galindo et al. 2011; Hamilton et al. 2012), but challenging to verify at the individual species level (Golden et al. 2014; Turvey et al. 2015). However, our study is unique because we compared local knowledge to three different external datasets and found that in all cases, fisher interviews were the most cost-effective method to generate spatial data.
Despite its great capacity to support conservation needs as in our study and others, (Thornton and Scheer 2012), one limitation of local knowledge is the lack of species-specific information, especially for hard-to-distinguish species (Turvey et al. 2014). Natural science methods (such as fishery research, scientific diving) or citizen science programs tend to be focused on species-level data, but take considerable amount of time and money to generate (Dickinson et al. 2010), as seen with our results. Local stakeholders may not initially be focused on gathering species level data (i.e., unaware of species differences or concentrating on other tasks) or may possess a folk taxonomy that distinguishes only amongst species that are commercially important (Beaudreau et al. 2011). However training stakeholders such as fishers to provide species level data forms the basis for many fishery dependent data collection methods (Stanley and Wilson 1990; Morgan and Burgess 2005). Indeed both commercial and small-scale fishers in Thailand record information on exploited species for Department of Fisheries research activities (R.Phoonsawat, pers.comm), and it may be feasible to train these fishers to identify and record seahorse species. While training activities are underway, genus level data provided by local knowledge can still provide a basis for modeling species distributions that can be further refined once additional data become available (Van Strien et al. 2013).
Our research comparing commercial and small-scale fisher knowledge differs from the usual practice of drawing mainly on small-scale fisher knowledge (Lunn and Dearden 2006; Golden et al. 2014). In many countries data from commercial fisheries are generated from logbooks, landings, or vessel monitoring systems, and rarely come from commercial fishers themselves but see (Hall et al. 2009; Shepperson et al. 2014; Turner et al. 2015). Indeed local knowledge in a fishery context generally discusses knowledge from small-scale fishers (Anuchiracheeva et al. 2003; Lunn and Dearden 2006). In terms of comparison to external datasets, our commercial fishers had sufficient knowledge to compare with fishery independent research as seen in other work (Hall et al. 2009; Shepperson et al. 2014). By contrast, we found small-scale fishers had more localized knowledge such as in Anuchiracheeva et al. 2003; Lunn and Dearden 2006 and covered similar spatial scales to diver generated datasets.
In contrast to most research, our results show that comparing local knowledge to governmental research trawls decreased the uncertainty associated with both datasets (Lewis et al. 2009; Hamilton et al. 2012). Commercial trawl captains provided a more complete spatial picture of our data-poor fishes whereas the DoF research trawls were more clearly quantitative, providing individual species information. Despite the increased length of time to generate the DoF dataset, and its high expense, one area with extremely high catch despite a smaller sampling effort was identified. Discrepancies in spatial or temporal data collection between fisher knowledge and fishery research can increase the uncertainty in both datasets if the conclusions drawn are not comparable (Usher 2000; Lewis et al. 2009). Such uncertainties have led to controversies surrounding the use of both local knowledge and scientific data to inform management (Usher 2000; Bohensky and Maru 2011). These controversies highlight the importance of knowledge integration and assessment criteria to ensure conservation can proceed in the face of uncertainty (Bohensky and Maru 2011; Hamilton et al. 2012).
Our comparison of diver generated data confirms that citizen science volunteers can provide data similar to scientists about species-level occurrence (Crall et al. 2011). If citizen science programs are well established, then informing the creation of a management plan with such data may be acceptable because of a large sample sizes or identification of appropriate analytical techniques (Crall et al. 2011; Van Strien et al. 2013). The iSeahorse.org program, a global citizen science initiative and database was launched in October 2013. Therefore when Thailand underwent the initial CITES enforcement process, the option to inform management was unavailable (UNEP-WCMC 2012). Citizen science programs are expensive (Dickinson et al. 2010). We coupled our iSeahorse citizen science activities with our scientific diving surveys to offset outreach costs, but it was still expensive. Two years of outreach efforts yielded a substantial amount of sightings (n = 95) considering our scientific dataset (n = 56) for the same area, over the same time period, failed to produce an equal number of sightings. However, compared to more established citizen science programs (e.g., ebird, 3.1 million bird sightings in North America for March 2012) it was a high cost with minimal return.
Five inherent biases may have influenced the complementarity of our datasets. First, by selecting to interview captains, we may have missed respondents with species level knowledge of seahorses because in some Thai fisheries (e.g., trawl fisheries), crew sort the fish and commonly set aside seahorses to sell in port. Second, there were large amounts of spatial and temporal variation in seahorse occurrence from the government research trawls, lending uncertainty to extrapolating seahorse locations at broad spatial scales. Third, challenges with detection of rare and cryptic species most likely biased scientific diving surveys, as documented in Aylesworth et al. (2015). Fourth, both the scientific diving surveys and citizen science contributions are biased towards shallow depths because of recreational safe diving limits. Lastly, there may be quality control issues with species identification in our citizen science contributions, as documented in other research (Dickinson et al. 2010). However the iSeahorse.org program allows participants the opportunity to request assistance with identification or report seahorse species as ‘unknown,’ which may have helped to minimize this uncertainty.
From our research we can deduce that various types of data gathering are beneficial for different management and conservation objectives, as discussed in other studies (Castellanos-Galindo et al. 2011; Beaudreau and Levin 2014). If you need to design and implement a species-specific management program (NOAA 2014), our research suggests that local fisher knowledge or citizen scientists can provide an informed starting point for determining species distribution. Such information can be refined to a species level at a later time through more advanced methods (e.g., scientific or fisheries independent surveys). If your objective is to identify key habitat areas as part of large scale conservation planning (NOAA 2014), then our research highlights that fisher knowledge is the most expeditious method to do so. If the management objective is to identify species threats, than our research, similar to others (Thornton and Scheer 2012), supports that stakeholder (fisher) interviews are the most cost effective data collection method. In our case, information about threats at the genus level was relevant to all species, and highlighted where current management and conservation efforts should be focused (e.g., specific gear types). While we found that fisher interviews enabled us to identify a path forward with management action, we recognize that they may not be an appropriate starting point in all circumstances because of perception bias, shifting baselines or telling researchers what they want to hear (Stake et al. 2005; Panjarat and Bennett 2012).
Taking the specific example of implementing CITES for seahorses in Thailand, our work exemplifies how spatial data can be used in the various stages of treaty implementation (CITES 2012; UNEP-WCMC 2012). CITES focus is to ensure that export levels do not harm local populations. To implement CITES successfully, Thailand needed to ensure the sustainability of wild populations that initially it could not locate. In this case, had local knowledge been available, Thailand could have prioritized gathering species-specific data in these locations. Since such information was unavailable, Thailand underwent the CITES enforcement process, where recommendations were given regarding actions to take to ensure sustainability (CITES 2012). One of these recommendations included undertaking studies to provide evidence on seahorse spatial abundance and use this information to consider spatial area restrictions for fisheries (CITES 2012). Our fisher knowledge provided spatial data at the genus level for where seahorses could be found, and also where fishing effort occurred. We are currently working to identify where management overlaps to evaluate if current management measures are appropriate to address the threats to local seahorse populations. Evident from our experience, is the importance of spatial data on distribution, threats and management to supporting CITES implementation for parties trading in Appendix II species.
Meeting international conservation actions are typically implemented by national governments, but to ensure effectiveness requires both top down and bottom up approaches (Moller and Berkes 2004; UNEP-WCMC 2012). With the example of seahorses in Thailand, the duty of CITES implementation for marine fishes lies within the DoF. Currently Thailand is working to create a monitoring and adaptive management plan for seahorses based largely on the spatial data related to distribution, threats and current management. If new management measures are enacted, previous studies in Thailand suggest their success may be linked to consultation with commercial and small-scale fishers (Lunn and Dearden 2006; Panjarat and Bennett 2012). In these instances fishers disagreed with the timing of closed seasons, doubted government enforcement capabilities, or were unaware of management measures, leading to lower compliance and support for restricted areas (Panjarat and Bennett 2012; Bennett and Dearden 2014). A key component for management effectiveness is promoting bottom up participation in natural resource governance (Moller and Berkes 2004; Berkes 2009), and encouraging expeditious data collection from resource users provides an opportunity for local input into the management process.
Our study supports the use of local knowledge to enable managers to act in the spirit of adaptive management for data-poor species (Walters and Holling 1990; Johannes 1998). Such knowledge can later be refined once more advanced data become available. With continued species declines both on land and in the ocean (McCauley et al. 2015), natural resource managers cannot afford to wait for the perfect dataset because the costs of inaction are high (Johannes 1998, 2000). The use of spatial data to evaluate conservation outcomes is growing in application for both international and national environmental agreements (Lorenz 1995; Lorenzen et al. 2010). Indeed spatial data are critical to achieve conservation and management objectives such as monitoring populations, protecting habitats or regulating extractive activities (Reid 2007; NOAA 2014). We encourage others involved at the interface of science, management and policy to consider gathering local knowledge prior to establishing initial management plans, prioritizing areas for conservation, and evaluating management effectiveness.
This is a contribution from Project Seahorse. The authors would like to thank the National Research Council of Thailand (Permit No. 0002/1306), Thailand Department of Fisheries, Phang-nga Provincial Marine Fisheries Station, Praulai Nootmorn, Tse-Lynn Loh, Sarah Foster, Sarah Harper, and Jennifer Selgrath. We are grateful for the support from numerous dive operators, fishers, and community groups who facilitated our search for seahorses.
This work was funded by the Ocean Park Conservation Foundation of Hong Kong, Riverbanks Zoo and Garden Conservation Fund, the Explorer’s Club Exploration Fund, SciFund Challenge, Bottom Billion Fieldwork Fund, FBR Capital Investments, John G. Shedd Aquarium, Guylian Chocolates and an anonymous donor.
Compliance with ethical standards
Conflict of interest
The authors declare they have no conflict of interest.
This research was conducted in accordance with UBC Animal (Permit No. A12-0288) and Human (Permit No. H12-02731) Ethics protocols. All participants interviewed gave an informed consent to participate in this research as per UBC Human Ethics protocols.